6 months to live for open models
5 hours ago
- #Policy Challenges
- #Open Source AI
- #AI Regulation
- The most serious test to date for open source AI viability is happening now due to policy actions that could marginalize open models.
- White House discussions may lead to new executive orders impacting open models, potentially starting with Chinese-origin models and government uses.
- Open models lack a central economic champion to defend against regulatory actions, unlike closed models with strong lobbying.
- Likely policy action includes banning or delaying open-weights models above capability levels like GPT 5.5 within the next 6 months.
- Two key policy discussions affecting open models are distillation (regulatory capture by companies like Anthropic) and frontier capabilities.
- Anthropic's campaign against Chinese models is seen as regulatory capture, aiming to ban competitors and secure economic advantages.
- Distillation debates focus on risks like Chinese labs distilling cybersecurity capabilities, but APIs are not magically secure and have vulnerabilities.
- A ban on open models could demolish the emerging open model economy in the US, including inference and fine-tuning companies.
- Handling frontier open-weight models like Mythos requires careful policy; a flat ban may be ineffective if not globally enforced and could isolate the US.
- Open models increase safety through broad access and understanding; stopping open-source progress requires global agreement, which is currently lacking.
- Short-term solutions include a US company releasing a capable open model to shift focus from China or building a coalition to lobby for safe open model rollout.